Category Archives: OS

CLI, OS

Using ProxMox as a NAS

9 Comments

Lately I’ve been very unhappy with latest FreeBSD causing reboots randomly during disk resilvering. I simply cannot tolerate random reboots of my fileserver. This fact combined with the migration of OpenZFS to the ZFS on Linux code base means it’s time for me to move from a FreeBSD based ZFS NAS to a Linux-based one.

Sadly there aren’t many options in this space yet. I wanted something where basic tasks were taken care of, like what FreeNAS does, but also supports ZFS. The solution I settled on was ProxMox, which is a hypervisor, but it also has ZFS support.

The biggest drawback of ProxMox vs FreeNAS is the GUI. There are some disk-related GUI options in ProxMox, but mostly it’s VM focused. Thus, I had to configure my required services via CLI.

Following are the settings I used when I configured my NAS to run ProxMox.

Repo setup

If you don’t want to pay for a proxmox license, change the PVE enterprise repository to the free version by modifying /etc/apt/sources.list.d/pve-enterprise.list to the following:

deb http://download.proxmox.com/debian/pve buster pve-no-subscription

Then run at apt update & apt upgrade.

Email alerts

Postfix configuration

Edit /etc/postfix/main.cf and tweak your mail server config as needed (relayhost). Restart postfix after editing:

systemctl restart postfix

Forward mail for root to your own email

Edit /etc/aliases and add an alias for root to forward to your desired e-mail address. Add this line:

root: YOUR_EMAIL_ADDRESS

Afterward run:

newaliases

ZFS configuration

Pool Import

Import the pool using the zpool import -f command (-f to force import despite having been active in a different system)

zpool import -f

By default they’re imported into the main root directory (/). If you want to have them go to /mnt, use the zfs set mountpoint command:

zfs set mountpoint=/mnt/

Monitoring

Install and configure zfs-zed 

apt install zfs-zed

Modify /etc/zfs/zed.d/zed.rc and uncomment ZED_EMAIL_ADDR, ZED_EMAIL_PROG, and ZED_EMAIL_OPTS. Edit them to suit your needs (default values work fine, they just need to be uncommented.) Optionally uncomment ZED_NOTIFY_VERBOSE and change to 1 if you want more verbose notices like what FreeNAS does (scrub notifications, for example.)

After modifying /etc/zfs/zed.d/zed.rc, restart zed:

systemctl restart zfs-zed

Scrubbing

By default ProxMox scrubs each of your datasets on the second Sunday of every month. This cron job is located in /etc/cron.d/zfsutils-linux. Modify to your liking.

Snapshot & Replication

There are many different snapshot & replication scripts out there. I landed on Sanoid. Thanks to SvennD for helping me grasp how to get it working.

Install sanoid :

#Install necessary packages
apt install debhelper libcapture-tiny-perl libconfig-inifiles-perl pv lzop mbuffer git
# Clone repo, build deb, install
git clone https://github.com/jimsalterjrs/sanoid.git cd sanoid
ln -s packages/debian . 
dpkg-buildpackage -uc -us 
apt install ../sanoid_*_all.deb

Snapshots

Edit /etc/sanoid/sanoid.conf with a backup and retention schedule for each of your datasets. Example taken from sanoid documentation:

[data/home]
	use_template = production
[data/images]
	use_template = production
	recursive = yes
	process_children_only = yes
[data/images/win7]
	hourly = 4

#############################
# templates below this line #
#############################

[template_production]
        frequently = 0
        hourly = 36
        daily = 30
        monthly = 3
        yearly = 0
        autosnap = yes
        autoprune = yes

Once sanoid.conf is to your liking, create a cron job to launch sanoid every hour (sanoid determines whether any action is needed when executed.) 

crontab -e
#Add this line, save and exit
0 * * * * /usr/sbin/sanoid --cron

Replication

syncoid (part of sanoid) easily replicates snapshots. The syntax is pretty straightforward: 

syncoid <source> <destination> -r 
#-r means recursive and is optional

For remote locations specify a username@ before the ip/hostname, then a colon and the dataset name, for example:

syncoid root@10.0.0.1:sourceDataset localDataset -r

You can even have a remote source go to a different remote destination, which is pretty neat.

Other syncoid options of interest:

--debug  #for seeing everything happening, useful for logging
--exclude  #Regular expression to exclude certain datasets
--src-bwlimit   #Set an upload limit so you don't saturate your bandwidth
--quiet #don't output anything unless it's an error

Automate synchronization by placing the same syncoid command into a cronjob:

0 */4 * * * /usr/sbin/syncoid --exclude=bigdataset1 --source-bwlimit=1M --recursive pool/data root@192.168.1.100:pool/data
#if you don't want status emails when the cron job runs, add --quiet

NFS

Install the nfs-kernel-server package and specify your NFS exports in /etc/exports.

apt install nfs-kernel-server portmap

Example /etc/exports :

/mnt/example/DIR1 192.168.0.0/16(rw,sync,all_squash,anonuid=0,anongid=0)

Restart nfs-server after modifying your exports:

systemctl restart nfs-server

Samba

Install samba, configure /etc/samba/smb.conf, and add users.

apt install samba
systemctl enable smbd

/etc/samba/smb.conf syntax is fairly straightforward. See the samba documentation for more information. Example share configuration:

[exampleshare]
comment = Example share
path = /mnt/example
valid users = user1 user2
writable = yes

Add users to the system itself with the adduser command:

adduser user1

Add those same users to samba with the smbpasswd -a command. Example:

smbpasswd -a user1

Restart samba after making changes:

systemctl restart smbd

SMART monitoring

Taken from https://pve.proxmox.com/wiki/Disk_Health_Monitoring:

By default, smartmontools daemon smartd is active and enabled, and scans the disks under /dev/sdX and /dev/hdX every 30 minutes for errors and warnings, and sends an e-mail to root if it detects a problem.

Edit the file /etc/smartd.conf to suit your needs. You can specify/exclude devices, smart attributes, etc there. See here for more information. Restart the smartd service after modifying.

UPS monitoring

apc-upsd was easiest for me to configure, so I went with it. Thanks to this blog for giving me the information to get started.

First, install apcupsd:

apt install apcupsd apcupsd-doc

As soon as it was installed my console kept getting spammed about IRQ issues. To stop these errors I stopped the apcupsd daemon:

 systemctl stop apcupsd

Now modify /etc/apcupsd/apcupssd.conf to suit your needs. The section I added for my CyberPower OR2200LCDRT2U was simply:

UPSTYPE usb
DEVICE

Then modify /etc/default/apcupsd to specify it’s configured:

#/etc/default/apcupsd
ISCONFIGURED=yes

After configuring, you can restart the apcupsd service

systemctl start apcupsd

To check the status of your UPS, you can run the apcaccess status command:

/sbin/apcaccess status

Log monitoring

Install Logwatch to monitor system events. Here is a good primer on all of Logwatch’s options. 

apt install logwatch

Modify /usr/share/logwatch/default.conf/logwatch.conf to suit your needs. By default it runs daily (defined in /etc/cron.daily/00logwatch). I added the following lines for my config to filter out unwanted information:

Service = "-zz-disk_space"
Service = "-postfix"
Service = "vsmartd"
Service = "-zz-lm_sensors"

Manually run logwatch to get a preview of what you’ll see:

logwatch --range today --mailto YOUR_EMAIL_ADDRESS

UPDATE 8/29/2020
I discovered additional tweaking to logwatch to get it exactly how I like it (thanks to this post and this one at serverfault.)

Defaults for monitored services are located in /usr/share/logwatch/default.conf/services/

You can copy this default file to /etc/logwatch/conf/services/<filename.conf> and then modify the service as needed. In my case I wanted to ignore logins for a particular user from a particular machine. This can be done by copying & editing sshd.conf and adding the following:

# Ignore these hosts
*Remove = 192.168.100.1
*Remove = X.Y.123.123
# Ignore these usernames
*Remove = testuser
# Ignore other noise. Note that we need to escape the ()
*Remove = "pam_succeed_if\(sshd:auth\): error retrieving information about user netscan.*

Troubleshooting

ZFS-ZED not sending email

If ZED isn’t sending emails it’s likely due to an error in the config. For some reason default values still need to be uncommented for zed to work, even if left unaltered. Thanks to this post for the info.

Samba share access denied

If you get access denied when trying to write to a SMB share, double check the file permissions on the server level. Execute chmod / chown as appropriate. Example:

chown user1 -R /mnt/example/user1
CLI, OS

zfs drive removal ‘part of active pool’ fix

Leave a comment

Occasionally I will manually offline a disk in my ZFS pool for one reason or another. Annoyingly I will sometimes get this error when I try to online that same disk back into the pool:

cannot online /dev/sda: cannot relabel '/dev/sda': unable to read disk capacity

The fix, thankfully, is fairly simple. Simply run the following command (make double sure you’re doing it on the correct device!)

sudo wipefs -a <DEVICE>

After I ran that command ZFS automatically picked the disk back and resilvered it into the pool.

Thanks to this superuser.com discussion for the advice!

CLI, Hardware, OS

Manually set monitor refresh rate with xrandr

Leave a comment

My recent distro hopping has found some distros don’t let you change your monitor’s refresh rate via the GUI. After a while I found this post which explains the syntax for xrandr to force the refresh rate of 100hz (adjust to your liking)

xrandr --output DP-0 --mode 3440x1440 --rate 100

Create an auto start item with this line of code in it to get this to happen automatically on startup.

CLI, OS

KeePassRPC incompatible with the current KeePass version

Leave a comment

I keep forgetting about this snag so I’ll document it. In Debian / Ubuntu distros, once you’ve added the PPA to have the latest version of KeePass installed, if you try to install the KeePassRPC plugin (Kee) it will tell you the version is incompatible, even though it is.

The following plugin is incompatible with the current KeePass version: /usr/lib/keepass2/Plugins/KeePassRPC.plgx

Have a look at the plugin's website for an appropriate version.

The fix, as found here, is to install the mono-complete package

sudo apt install mono-complete

Restart KeePass after installation. That’s it!

OS

Create local CentOS 7 Repo

Leave a comment

I’ve recently needed to create a local mirror of Cent7 packages. I followed the guide posted on techmint but also made a few tweaks to get it to work to my liking.

Create local repo mirror

  • Install necessary packages
    • sudo yum -y install epel-release nginx createrepo yum-utils moreutils
  • Create directories that will host your repo
    • sudo mkdir -p /usr/share/nginx/html/repos/{base,centosplus,extras,updates,epel}
  • Use the reposync tool to synchronize to those local directories (repeat for each directory, changing repoid= value to match)
    • reposync -g -l -d -m --repoid=base --newest-only --download-metadata --download_path=/usr/share/nginx/html/repos/
  • Use the createrepo tool to create repodata
    • base & epel have a group file, other repos do not.
    • For base & epel:
      • createrepo -g comps.xml /usr/share/nginx/html/repos/<FOLDER>
    • For the rest:
      • createrepo /usr/share/nginx/html/repos/<FOLDER>

Configure daily synchronization via cron

Copy this script to /etc/cron.daily/ and give it execute rights

#!/bin/bash
##specify all local repositories in a single variable
LOCAL_REPOS="base extras updates epel centosplus"
##a loop to update repos one at a time
for REPO in ${LOCAL_REPOS}; do
reposync -g -l -d -m --repoid=$REPO --newest-only --download-metadata --download_path=/usr/share/nginx/html/repos/Cent7/
if [[ $REPO = 'base' || $REPO = 'epel' ]]; then
        createrepo -g comps.xml /usr/share/nginx/html/repos/Cent7/$REPO/
else
        createrepo /usr/share/nginx/html/repos/Cent7/$REPO/
fi
done
chmod 755 /etc/cron.daily/<script_name>

E-mails from cron became annoying. I wanted to only get e-mailed on error. The solution is to use chronic

Modify /etc/anacrontab to add “chronic” between nice and run-parts

1 5 cron.daily nice chronic run-parts /etc/cron.daily
7 25 cron.weekly nice chronic run-parts /etc/cron.weekly
@monthly 45 cron.monthly nice chronic run-parts /etc/cron.monthly

Success.

Update 4/25/19 I encountered an issue while trying to use repsync to mirror the remi repo.

warning: /usr/share/nginx/html/repos/Cent7/remi/remi/aspell-nl-0.50-1.el7.remi.x86_64.rpm: Header V4 DSA/SHA1 Signature, key ID 00f97f56: NOKEY

I found out from here that it means you need to manually import the package’s key into the RPMDB like so

sudo rpm --import /etc/pki/rpm-gpg/RPM-GPG-KEY-remi

OS, Virtualization

Proxmox first VM boot delay workaround

2 Comments

My home lab has an NFS server for storage and a proxmox hypervisor connecting to it. If the power ever goes out for more than my UPS can handle, startup is a bit of a mess. My ProxMox server boots up much faster than my NFS server, so the result is no VMs start automatically (due to storage being unavailable) and I have to manually go in and start everything.

I found this bug report from 2015 which frustratingly doesn’t appear to have any traction to it. Ideally I could just tell the first VM to wait 5 minutes before turning on, and then trigger all the other VMs to turn on once the first one is up, but the devs don’t seem to want to address that issue. So, I got creative.

My solution was to alter the grub menu timeout before booting ProxMox. Simple but effective.

Edit /etc/default/grub and modify GRUB_TIMEOUT

#modify GRUB_TIMEOUT to your liking
GRUB_TIMEOUT=300

Then simply run update-grub

update-grub

Now my proxmox server waits 5 minutes before even booting the OS, by which time the NAS should be up and running. No more manual turning on of VMs after a power outage.

CLI, OS, Virtualization

Fix Proxmox swapping issue

1 Comment

I recently had an issue with one of my Proxmox hosts where it would max out all swap and slow down to a crawl despite having plenty of physical memory free. After digging and tweaking, I found this post which directed to set the kernel swappiness setting to 0. More reading suggested I should set it to 1, which is what I did.

Append to /etc/sysctl.conf:

#Fix excessive swap usage
vm.swappiness = 1

Apply settings with:

sysctl --system

This did the trick for me.

CLI, Hardware, OS

Rasbperry Pi as a dashboard computer

Leave a comment

Here are my raw, unpolished notes on how I set up a raspberry pi to serve as a dashboard display:

Use Raspbian OS

Autostart Chrome in kiosk mode

Eliminate Chrome crash bubble thanks to this post

mkdir -p ~/.config/lxsession/LXDE-pi/
nano ~/.config/lxsession/LXDE-pi/autostart

Add this line:
@chromium-browser --kiosk --app=<URL>

Mouse removal

sudo apt-get install unclutter

in ~/.config/lxsession/LXDE-pi/autostart add

@unclutter -idle 5

Disable screen blank:

in /etc/lightdm/lightdm.conf add

[SeatDefaults]
xserver-command=X -s 0 -dpms

Open up SSH & VNC

Pi / Preferences / Raspberry Pi Configuration: Interfaces tab

SSH: Enable
VNC: Enable

Increase swap file

sudo nano /etc/dphys-swapfile
CONF_SWAPSIZE=2048M

Configure NTP

sudo apt-get install openntpd ntpdate
sudo systemctl enable openntpd
sudo ntpdate <IP of NTP server>

edit /etc/openntpd/ntpd.conf and modify servers lines to fit your NTP server

Disable overscan

Pi / Preferences / Raspberry Pi Configuration: System tab
Overscan: Disable

CLI, OS

Recover files from ZFS snapshot of ProxMox VM

Leave a comment

I recently needed to restore a specific file from one of my ProxMox VMs that had been deleted. I didn’t want to roll back the entire VM from a previous snapshot – I just wanted a single file from the snapshot. My snapshots are handled via ZFS using FreeNAS.

Since my VM was CentOS 7 it uses XFS, which made things a bit more difficult. I couldn’t find a way to crash-mount a read-only XFS snapshot – it simply resufed to mount, so I had to make everything read/write. Below is the process I used to recover my file:

On the FreeNAS server, find the snapshot you wish to clone:

sudo zfs list -t snapshot -o name -s creation -r DATASET_NAME

Next, clone the snapshot

sudo zfs clone SNAPSHOT_NAME CLONED_SNAPSHOT_NAME

Next, on a Linux box, use SSHFS to mount the snapshot:

mkdir Snapshot
sshfs -o allow_other user@freenas:/mnt/CLONED_SNAPSHOT_NAME Snapshot/

Now create a read/write loopback device following instructions found here:

sudo -i #easy lazy way to get past permissions issues
cd /path/to/Snapshot/folder/created/above
losetup -P -f VM_DISK_FILENAME.raw
losetup 
#Take note of output, it's likely set to /dev/loop0 unless you have other loopbacks

Note if your VM files are not in RAW format, extra steps will need to be taken in order to convert it to RAW format.

Now we have an SSH-mounted loopback device ready for mounting. Things are complicated if your VM uses LVM, which mine does (CentOS 7). Once the loopback device is set, lvscan should see the image’s logical volumes. Make the desired volume active

sudo lvscan
sudo lvchange -ay /dev/VG_NAME/LV_NAME

Now you can mount your volume:

mkdir Restore
mount /dev/VG_NAME/LV_NAME Restore/

Note: for XFS you must have read/write capability on the loopback device for this to work.

When you’re done, do your steps in reverse to unmount the snaspshot:

#Unmount snapshot
umount Restore
#Deactivate LVM
lvchange -an /dev/VG_NAME/LV_NAME
Remove loopback device
losetup -d /dev/loop0 #or whatever the loopback device was
#Unmount SSHfs mount to ZFS server
umount Snapshot

Finally, on the ZFS server, delete the snapshot:

sudo zfs destroy CLONED_SNAPSHOT_NAME

Troubleshooting

When I tried to mount the LVM partition at this point I got this error message:

mount: /dev/mapper/centos_plexlocal-root: can't read superblock

It ended up being because I was accidentally creating a read-only loopback device. I destroy the loopback device and re-created with write support and all was well.